A COMPARISON OF DIVIDEND, CASH FLOW, AND EARNINGS APPROACHES TO EQUITY VALUATION

Stephen H. Penman Walter A. Haas School of Business University of California, Berkeley Berkeley, CA 94720 (510) 642-2588 and Theodore Sougiannis College of Commerce and Business Administration University of Illinois at Urbana-Champaign Champaign, IL 61820 (217) 244-0555

January, 1995 Revision: April, 1996

We thank Pat O'Brien, Jim Ohlson, Mike Oleson, Morton Pincus, Stephen Ryan, Jacob Thomas and Dave Ziebart for comments.

ABSTRACT

Standard formulas for valuing equities require prediction of payoffs "to infinity" for going concerns but a practical analysis requires that they be predicted over finite horizons.

This truncation

inevitably involves (often troublesome) "terminal value" calculations. This paper contrasts dividend discount techniques, discounted cash flow analysis, and techniques based on accrual earnings when applied to a finite-horizon valuation.

Valuations based on average ex post payoffs

over various horizons, with and without terminal value calculations, are compared with (ex ante) market prices to give an indication of the error introduced by each technique in truncating the horizon.

Comparisons of

these errors show that accrual earnings techniques dominate free cash flow and dividend discounting approaches.

Further, the relevant

accounting features of each technique are identified and the source of the accounting that makes it less than ideal for finite horizon analysis (and for which it requires a correction) are discovered.

Conditions

where a given technique requires particularly long forecasting horizons are identified and the performance of the alternative techniques under those conditions is examined.

A COMPARISON OF DIVIDEND, CASH FLOW, AND EARNINGS APPROACHES TO EQUITY VALUATION

The calculation of equity value is typically characterized as a projection of future payoffs and a transformation of those payoffs into a present value (price).

A good deal of research on pricing models has

focused on the specification of risk for the reduction of the payoffs to present value but little attention has been given to the specification of payoffs.

It is noncontroversial that equity price is based on future

dividends to shareholders but it is well-recognized that dividend discounting techniques have practical problems.

A popular alternative--

discounted cash flow analysis--targets future "free cash flows" instead. Analysts also discuss equity values in terms of forecasted earnings and the classical "residual income" formula directs how to calculate price from forecasted earnings and book values.

It is surprising that, given

the many prescriptions in valuation books and their common use in practice, there is little empirical evaluation of these alternatives.1 This paper conducts an empirical examination of valuation techniques with a focus on a practical issue.

Dividend, cash flow and

earnings approaches are equivalent when the respective payoffs are predicted "to infinity," but practical analysis requires prediction over finite horizons. known.

The problems this presents for going concerns are well

In the dividend discount approach, forecasted dividends over the

immediate future are often not related to value so the forecast period has to be long or an (often questionable) terminal value calculation made at some shorter horizon.

Alternative techniques forecast "more

fundamental" attributes within the firm instead of distributions from

2 the firm.

However this substitution solves the practical problem only

if it brings the future forward in time relative to predicted dividends, and these techniques frequently require terminal value corrections also. In discounted cash flow (DCF) analysis the terminal value often has considerable weight in the calculation but its determination is sometimes ad hoc or requires assumptions regarding free cash flows beyond the horizon.

Techniques based on forecasted earnings make the

claim (implicitly) that accrual adjustments to cash flows bring the future forward relative to cash flow analysis, but this claim has not been substantiated in a valuation context. The paper assesses how the various techniques perform in finite horizon analysis.

What techniques work best for projections over one,

two, five, eight year horizons and under what circumstances?

A

particular focus is the question of whether the projection of accounting earnings facilitates finite horizon analysis better than DCF analysis. Analysts typically forecast earnings but, for valuation purposes, should these be transformed to free cash flows?

In classroom exercises

students are instructed to adjust forecasted earnings for the accruals to "get back to the cash flows."

This is rationalized by ideas that

cash flows are "real" and the accounting introduces distortions, but is the exercise warranted? The valuation techniques are evaluated by comparing actual traded prices with intrinsic values calculated, as prescribed by the techniques, from subsequent payoff realizations.

Ideally one would

calculate intrinsic values from unbiased ex ante payoffs but, as forecasts are not observable for all payoffs, intrinsic values are

3 calculated from average ex post payoffs.2

Firm realizations are

averaged in portfolios and portfolio values are then pooled over time to average out the unpredictable component of ex post realizations. Intrinsic values calculated from these realizations are compared with actual prices to yield ex post valuation errors and, if average realizations represent ex ante expectations, estimates of ex ante errors on which the techniques are compared.

Both mean errors and the

variation of errors are considered as performance metrics.

This

comparison is made under the assumption that, on average, actual market prices with which calculated intrinsic values are compared are efficient at the portfolio level with respect to information that projects the payoffs. Valuation techniques are characterized as pro forma accounting methods with different rules for recognizing payoffs, and their relevant features are identified within a framework that expresses them as special cases of a generic accounting model.

This framework refers to

the reconciliation of the infinite horizon cash flow and accrual accounting models in Feltham and Ohlson (1995) and the finite-horizon synthesis in Penman (1996).

It establishes conditions where each

technique provides a valuation without error, with and without terminal values, and identifies when (seemingly different) calculations yield the same valuation.

In particular, it demonstrates that DCF techniques with

"operating income" specified in the terminal value are identical to models that specify accrual earnings as the payoff.

Hence the

comparison of DCF techniques with accrual accounting residual income techniques amounts to comparing different calculations of the terminal

4 value in DCF analysis.

This brings the focus to the critical practical

problem, the determination of terminal values. This framework dictates the construction of the empirical tests. Conditions where a particular technique is ideal (for a finite-horizon analysis) are identified and the error metrics for the techniques are calculated over departures from this ideal.

Thus the aspect of the

technique's accounting that produces error is identified.

Then error

metrics for alternative techniques are calculated over the same conditions to assess improvement (or otherwise) that can be identified with the different accounting.

In this way we develop an appreciation

of how alternative accounting works for valuation purposes. The analysis quickly dismisses dividend discounting techniques as inappropriate for finite horizons.

It shows that techniques based on

GAAP earnings dominate those based on cash flows.

It demonstrates

explicitly that the accrual accounting involved in earnings techniques provides a correction to the discounted cash flow valuation.

This

involves the accounting for anticipated investment and the recognition of non-cash value changes.

It also compares discounted residual

earnings approaches and capitalized earnings approaches under a variety of conditions.

Finally, it identifies conditions where earnings

approaches, while dominating discounted cash flow techniques, do not perform particularly well over five to eight year horizons.

These are

associated with high price-to-earnings and extreme price-to-book firms. Section I describes the accounting involved in various valuation approaches.

Section II outlines valuation over finite-horizons,

identifies conditions where the techniques yield valuations without

5 error, and demonstrates some equivalences between techniques. Section III outlines the research design and the data sources, and Section IV presents the results.

I.

EQUITY VALUATION TECHNIQUES

A.

The Dividend Discount Approach The theory of finance describes equity valuation in terms of

expected future dividends.

Formally,

Pt =

∑ ρ E( ~d ) -τ

τ =1

t+τ

(DDM)

(1)

where Pt is the price of equity at time t, dt+ τ is net dividends paid at t+τ, ρ is one plus the discount rate (equity cost of capital), indicated as a constant, and E is an expectation conditional on information at time t.

Firm subscripts are understood.3

This dividend discount model

(DDM) targets the actual distributions to shareholders but, despite this appeal, its application in practice (over finite horizons) is viewed as problematic.

The formula requires the prediction of dividends to

infinity or to a liquidating dividend but the Miller and Modigliani (1961) dividend irrelevance proposition states that price is unrelated to the timing of expected payout prior to or after any finite horizon. So, for going concerns, targeted dividends to a finite horizon are uninformative about price unless policy ties the dividend to valuegenerating attributes.

This calls for the targeting of something "more

fundamental" than dividends.

6

B.

Generic Accounting Approaches In recognition of this so-called dividend conundrum, alternative

valuation approaches target attributes within the firm which are conjectured to capture value creating activities rather than the valueirrelevant payout activities.

The identification and tracking of

additions to value is an accounting system.

An accounting system that

periodically recognizes additions to value that are distinguished from distributions of value is expressed as:

Bt+τ = Bt+τ -1 + Xt+τ - d t+τ ,

for all τ.

(CSR)

(2)

In this "clean surplus relation," Bt+ τ is the measured stock

of value ("book value") at t+τ, Xt+ τ is the measured flow of added value ("earnings") from t+τ-1 to t+τ (calculated independently of dividends), and the dividends are negative for equity contributions.

It is well-

recognized (in Preinreich (1938), Edwards and Bell (1961) and Peasnell (1982), for example) that, solving for dt+ τ in the CSR equation and substituting into (1),

PTt ≡ B t +

∑ ρ E[ X~ - (ρ -1) ~B ] -τ

τ =1

t+τ

(3)

t+τ -1

approaches Pt in (1) at T→∞, given a convergence condition similar to that for the dividend discount formula.

The expression over which the

expectation is taken compares future flows to those projected by applying the discount rate to beginning-of-period stocks.

This equation

holds for all clean-surplus accounting principles and alternative

7 valuation techniques are distinguished by the identification of B and X and the rules for their measurement.

In this respect, a valuation

technique and a (pro forma) accounting system (for equity valuation) are the same thing.

C.

Accounting for Financial Activities and Discounted Cash Flow Analysis A common approach substitutes "free cash flows" for dividends as

the target of analysis (for example, in Rappaport (1986), Copeland, Koller, and Murrin (1990), Hackel and Livnat (1992) and Cornell (1993)). The standard derivation begins with the cash conservation equation (CCE):

Ct+τ - I t+τ ≡ d t+τ - F t+τ ,

all τ ,

(CCE)

(4)

where C is cash flow from operations, F is cash flow from non-equity financing activities, I is cash investment, and d is dividends net of equity contributions (as before).

Let FAt denote the present value of

future cash flows with respect to financing activities (net financial assets).

Then, solving CCE for dt+ τ and substituting into (1),

Pt =

∑ ρ E(~C τ =1

-τ w

t+τ

~I ) t+τ

+

FA t ,

(5)

where Ct+ τ - It+ τ is called "free cash flow" and FAt is usually indicated as negative (net debt) to reflect net borrowing rather than lending. The discount rate, ρw , is the weighted-average (unlevered) cost of capital, recognizing (as in Modigliani and Miller (1958)) that the

8 operation's cost of capital is independent of financing. Feltham and Ohlson (1995) demonstrate that this expression can also be derived from the stocks and flows equation (CSR). special case of (3) with a particular accounting.

Thus (5) is a

This accounting

identifies Bt+ τ ≡ FAt+ τ and Xt+ τ ≡ Ct+ τ - It+ τ + it+ τ , all τ, where it+ τ is cash interest on financial assets which, with principal flows, is part of Ft+ τ and which is negative for net debt.

Thus the clean surplus equation,

FAt+ τ = FAt+ τ -1 + Ct+ τ - It+ τ + it+ τ - dt+ τ , describes an accounting system that tracks financial assets (or debt).

Free cash flows are invested in

financial assets (or reduce debt) and dividends are paid out of financial assets.

This merely places the CCE flow equation on a stocks

and flows basis as the net addition to financial assets (net of interest) is equal to Ft+ τ , by CCE.

PTt = FA t +

The calculation in (3) becomes

~ ~ ~ ∑ ρ E (C- I+ i ) - (ρ -1)~FA -τ

t+τ

τ =1

t+τ -1

 .

(6)

* Replacing it+ τ with i t+τ such that

~ ∑ ρ E(i -τ

τ =1

* t+τ

)=

then

PTt = FA t +

∑ ρ ( ρ -1)E(~FA -τ

∑ ρ E(~C τ =1

approaches Pt in (5) and (1) as T→∞.

-τ w

t+τ

~I t+τ

),

(7)

t+τ -1

τ =1

),

(DCFM)

(8)

Condition (7) requires that

interest be accounted for on accrual basis independent of the cash coupon (the "effective interest" method) and correspondingly FAt+ τ is, in

9 expectation, at present value (market value) for all τ≥0.

We refer to

(8) as the discounted cash flow model, DCFM. This is an accounting system that tracks financial activities. The book value of equity is the value of the bonds and the technique for the valuation of bonds is appropriated for the valuation of equity. Correspondingly, the targeted flow reflects financing flows.

For a firm

with no financial assets or debt (an "all equity" firm, for example), free cash flow, Ct+ τ - It+ τ ≡ dt+ τ , by CCE, and hence the target is the same as in the dividend discount formula with the same problems induced by dividend irrelevance.

The clean-surplus system that is nominated to

distinguish value added activities from dividend activities degenerates to tracking dividends.

For a firm with debt financing, Ct+ τ - It+ τ ≡

dt+ τ - Ft+ τ , but the adjustment to dividends for financing flows introduces a zero net present value attribute which is irrelevant to value (Modigliani and Miller (1958)).

Value is deemed to be created by

operational activities but this technique targets financing stocks and flows rather than operating stocks and flows.

As Ct+ τ applies to

operations, it is the negative treatment of investment in the free cash flow measure of value added that produces this.

10

D.

Accounting for Financial and Operating Activities and Earnings Approaches to Valuation Feltham and Ohlson (1995) characterize clean-surplus accounting

systems that incorporate operating activities. FAt+ τ + OAt+ τ .

Identify Bt+ τ ≡

OAt+ τ is a measure of operating assets (net of operating

liabilities) which are accounted for as OAt+ τ = OAt+ τ -1 + It+ τ + oat+ τ where By CSR, Xt+ τ = ∆(FAt+ τ + OAt+ τ ) + dt+ τ

oat+ τ is measured operating accruals.

* (where ∆ indicates changes) and thus, as ∆FAt+ τ = Ct+ τ - It+ τ + i t+τ - dt+ τ , * as before, Xt+ τ = Ct+ τ + i t+τ + oat+ τ , where Ct+ τ + oat+ τ ≡ OIt+ τ is commonly

referred to as operating income.

Financial assets are booked at present *

value, as before, and thus interest is accrued into i t+τ .

Investments

are booked as part of operating assets rather than part of the value added flow and, in addition, other non-cash flow values (like receivables) are recognized as value added in the accruals. U.S. GAAP bears a strong resemblance to this accounting. from (3),

[

∑ τ =1

- (ρ -1)(

~ FA t+τ -1

+

~ OA t+τ -1

Current

Accordingly,

)]

(RIM)

(9)

and, given the financial accrual condition in (7),

PTt = FA t + OA t +

~ ∑ ρ E[OI τ =1

-τ w

t+τ

~ - (ρ -1)OA

],

t+τ -1

The target in (9) is referred to as (accrual accounting) "residual income" and we refer to (9) as the residual income model (RIM).

(10)

11 Equation (10) reflects that financing is at zero net present value and therefore drops out.

The target, operating income less a charge against

operating assets, has been popularized as "Economic Value Added" by Stewart (1991).

E.

The Coca Cola Co. refers to it as "economic profit."

Accounting Approaches Involving Capitalization Ohlson (1995) shows that by iterating out flows from sequential

book values in (3) (with no further assumptions),

 ~ VTt ≡ ( ρ T -1 ) -1E ∑ X t+τ  τ =1

approaches Pt in (1) and (3) as T→∞.

+ ∑(ρ τ =1

T- τ

 -1) ~ d t+τ  

(CM)

(11)

This involves adjusting expected

earnings within the firm for earnings from reinvesting the dividends paid out and capitalizing the aggregated cum-dividend flow at the cost of capital.

It can be shown that

(

~ VTt = Bt + (ρ T -1 )-1 ∑ ρ T-τ E Xt+τ τ =1

- (ρ -1)~B

T so, for all T, V t is current book value plus the capitalized terminal

value of the expected residual income in (3) rather than its present value.

Like (3) it holds for all clean-surplus specifications of

X and B and the free cash flow and accrual accounting specifications are special cases.

Easton, Harris and Ohlson (1992) show that the

cum-dividend earnings (within the square parentheses), measured according to GAAP, are highly correlated with stock returns over five to ten year periods.

t+τ -1

)

12

II.

VALUATION OVER FINITE HORIZONS Clearly all specifications of X and B and both the discounting and

capitalization approaches produce the same valuation when attributes are projected "to infinity," and this equals the valuation for the infinitehorizon dividend discount formula.

The practical issue is what

specifications are appropriate for finite horizon forecasting and under what conditions. By iterating out dividends from successive X and B (by CSR), the generic calculation in (3) can be stated as

PTt =

~ ∑ ρ E(d -τ

t+τ

τ =1

)+

( ),

ρ -T E ~ Bt+T

(12)

that is, the present value of forecasted dividends to t+T plus the present value of the expected t+T stock.

As, for DCF analysis,

Bt+T ≡ FAt+T and for RIM, Bt+T ≡ FAt+T + OAt+T , the two valuations differ for a given horizon, t+T, by the present value of expected t+T operating assets, and are the same only when operating assets are projected to be liquidated (into financial assets). Further, the DDM in (1) for a finite t+T is expressed as

Pt =

~ ∑ ρ E(d -τ

τ =1

by the no-arbitrage condition.

t+τ

)+

( )

ρ -T E ~ Pt+T

(13)

Thus, for any specification of X and B,

T valuation is made without error (P t = Pt ) if

~ ) = 0 (by E(~ Pt+T - B t+T

T comparing (12) and (13)), and the error of P t is

ρ -T E (~ Pt+T - ~ Bt+T).

13 Accordingly, the DCF analysis will yield the correct valuation only if operating assets are to be liquidated into financial assets (measured at market value), and RIM will yield the correct valuation if expected t+T operating assets are at market value.

For the CM approach in (11),

T valuation without error (Vt = Pt ) occurs if

~ )E(~ Pt+T - B t+T

(Pt - Bt)

= 0,

that is when there is no expected change in the calculated premium to the horizon, and the error is given by the present value of the expected change in premium (Ohlson (1995)).

t

The zero error conditions for both

T

P

T and Vt have the feature that the accounting brings the future forward

in time such that forecasting to the horizon is sufficient for forecasting "to infinity."

T For P t the forecasted book value at t+T is

sufficient for subsequent flows (and for expected price at t+T) and for V T t

aggregated (cum-dividend) flows to t+T are sufficient for projecting

subsequent flows at the cost of capital. These zero error conditions are restrictive.

DCF analysis cannot

be used for firms with continuing operations and Ou and Penman (1995) show that neither condition is representative in the cross section with GAAP accounting over any "reasonable" horizon.

"Terminal value"

corrections are typically required, as recognized in practice. Penman (1996) provides a general model of finite-horizon valuation T T which includes P t and V t as special cases. t+T+NS

-

t+T+NS

) = Ks E(t+T -

t+T

If, for a horizon t+T, E(

) τ for all N>0 and a given S>0, then =1

 S   S  +ρ (ρS -Ks) -1 E  ∑ X~ t+T+ τ + ∑ ρS-τ -1 ~dt+T+τ - (Ks -1) B~t+T  -E ~Bt+T  τ =1 τ =1   -T

(

)

( )

(14)

14

provides the valuation, Pt , without error, and this valuation can be restated as

τ =1

 S S  (14a)  S  -1 +ρ ( ρ - Ks ) E ∑ X~ t+T+τ + ∑ ρS-τ -1 ~d t+T+τ -( ρs -1) ~Bt+T  . τ =1 τ =1 

(

-T

)

The expected changes in premiums that Ks projects are differences in cum-dividend flows relative to cum-dividend changes in value, by CSR, and thus (the constant) Ks captures projected errors in measuring value added, consistently applied.

This constant measurement error is

manifest in forecasted S-period expected residual income growing subsequent to t+T at the rate Ks -1, and accordingly can be inferred. The standard terminal value calculation based on perpetual growth of some attribute is of course consistent with this.

It sets S = 1 and

capitalizes at the rate ρ-K1 where K1 is the one period growth rate. The formulation here gives this an accounting measurement error interpretation, generalizes it as an S-period calculation, and points out that it is the forecasted growth in residual earnings rather than earnings that indicates Ks , the measurement error on which the terminal value is based.

T T T* P t combines P t and V t into a general valuation

T formula and P t = Pt is a special case when the last term in (14) is T zero and V t = Pt (another special case) when Ks = 1 and T = 0.

This formulation yields the generalized terminal value for the DDM.

As the last term in (14) gives the error,

in (13) is supplied:

~ ), then E(t+T ) E(~ Pt+T - B t+T

15

+

 ρ -T  ρ S -Ks 

(

) E ∑ X~ -1

τ =1

t+T+τ

+ ∑ (ρ τ =1

S-τ

~ ~  -1 d t+T+τ - (Ks -1) Bt+T   

)

(15)

= PT* t

(Penman (1996)).

This provides an umbrella over all other calculations:

the specification of X and B and the calculation of price according to (14) reduces to the question of the appropriate specification of the terminal value for the dividend discount model.

The specification of

attributes to be forecasted to the horizon is not important.

All

valuations can be expressed in terms of a cum-dividend terminal value for the DDM and it is this calculation that is the determining one. This umbrella identifies calculations that look different but are in fact the same.

To be less cumbersome, set S = Ks = 1 and so (15)

becomes

Pt =

(which equals

~ ∑ ρ E(d -τ

t+τ

τ =1

) + ρ [(ρ -1) -T

-1

E(X~

t+T+1

)]

(15a)

PT* in (14)). With the DCF specification, this is stated t τ =1

T

= ∑ ρ E( ) + τ =1

-τ

[

( )

∑

as

~ d t+τ

[

~ ~ ρ (ρ w -1) E (CI )t+T+1 -T

-1

and for the accrual accounting specification,

]

+ E(

~ FA t+T

)],

(15b)

16

τ =1

T

= ∑ ρ E( ) + ~ d t+τ

-τ

τ =1

[

( )

∑

and so for S > 1 and Ks > 1.

ρ

-T

[(ρ -1) w

]

-1

E(

)+(

~ OIt+T+1

~ FA t+T

(15c)

)]

Thus, given the premium (error) condition

under which (14) yields the price for the accrual accounting model, the DCF valuation will also yield the same price for the same horizon (only) if

~ ~ ~ E (C- I )t+T+1 = E(OI Further, Penman (1996) shows t+T+1) , and vice versa.

that the practice of specifying capitalized operating income as the terminal value calculation in DCF analysis such that,

~ ~ I) + ∑ ρ E (C-

Pt = FA t +

τ =1

-τ w

t+τ

ρ -Tw

[( ρ -1) E(OI~ )] -1

w

t+T+1

is equivalent to (15c), the accrual accounting calculation.

(15d)

In effect,

this is not cash flow analysis at all, but rather accrual accounting, and contrasts to the pure DCF analysis in (15b) which, stated in the form of (14a) for Ks = S = 1 (as is usual), is

[( ρ -1) E (C-~ ~I+ ~i ) - E(~FA )] ~ ~ I ) ], + ∑ ρ E(~C- ~I) + ρ [(ρ -1) E (C+

=

ρ -T

T

FA t

τ +1

-τ w

-1

*

t

t+T+1

-T

t+τ

with the accommodation for S > 1 and Ks > 1.

w

(16)

-1

t+T+1

As Ct+T+1 - It+T+1 ≡ dt+T+1 -

Ft+T+1 , this amounts to capitalizing financing flows that are forecasted

17 to be a constant in perpetuity.

Accordingly we examine accrual

accounting against the pure DCF analysis with the understanding that this can be stated as a comparison of the terminal value calculation for DCF analysis in (15d) with that in (16).4

III.

DATA AND RESEARCH DESIGN The empirical analysis compares valuations based on the DDM, DCFM,

RIM and CM over various horizons, with and without the terminal value calculations in (14).

Valuations at time t are calculated from

subsequent realizations of the X and B specified by the alternative models up to various t+T+1 and these are then compared with actual traded price at t. This design relies on assumptions required to infer ex-ante values from ex-post data.

We assume that (a) average realizations are equal to

their ex-ante rational expectations, and (b) observable market prices to which calculated intrinsic values are compared are efficient. Accordingly, the analysis is on portfolios of stocks observed over time with the aim of averaging out unexpected realizations and any market inefficiencies over firms and over time. We first evaluate the valuation methods over all conditions and then under various circumstances where the accounting may affect the horizon over which analysis is done.

The analysis over all conditions

is implemented by random assignment of firms to portfolios. conditional tests assign firms to portfolios on the basis of conditioning circumstances.

The

18 For the unconditional tests, firms are randomly assigned to 20 portfolios at the end of each year of the sample period, t = 1973-1990.

Arithmetic average portfolio values of the respective

accounting realizations are then calculated for each subsequent ten years (t+T, T=1,2,…,10) and ex post intrinsic values of common equity are calculated at the end of year t from these mean realizations according to the prescription of the relevant formula for each horizon, t+T.

The respective techniques are evaluated on (ex post) errors of

these values relative to observed price at the end of year t.

Mean

errors and the variation in errors are then calculated over all 18 years.5 The data used in this study are taken from the COMPUSTAT Annual and Research files which cover NYSE, AMEX, and NASDAQ firms.

The

combined files include non-surviving firms to the year of their termination.

The files cover the period 1973 to 1992.

Financial firms

(industry codes 6000-6499) are not included in the analysis.

The number

of firms available for each year (with prices, dividends, and accounting data for that year) range from 3544 in 1973 to 5642 in 1987, with an average of 4192 per year.

As there are no data after 1992, the number

of years in the calculations declines as the horizon increases.

For

ten-year horizons (T=10), there are 10 years (1973-82) and for T=1, there are 18 years (1973-90). The exercise raises a number of issues about the accounting for the attributes and these are addressed in Appendix A.

The cost of

capital determination is elusive and we applied a number of calculations.

For the equity cost of capital we used, alternatively:

19 the risk free rate (the 3-year T-Bill rate p.a.) for the relevant year plus an equity risk premium of 6% p.a. for all firms (approximately the historical equity premium reported in Ibbotson and Sinquefield (1983) at the beginning of the sample period); the cost of capital given by the CAPM using the same risk free rate and risk premium with betas estimated for each firm; and the cost of capital for the firm's industry based on the Fama and French (1994) three factor (beta, size and book-to-price) model.6

These all were updated each year.

for all firms in all years.

Finally, we used a 10% rate

We report results with CAPM estimates (and

the notation, ρ, will imply this) but little difference in results was observed with the calculations, and it will become apparent that reasonable risk adjustment cannot explain the results.

For discounting

or capitalizing operational flows, an unlevered cost of capital was calculated using standard techniques.7 The study is concerned with ex ante going-concern valuation but firms terminate ex post.

Appendix B describes how the calculations deal

with this to accommodate questions of survivorship.

20

IV. A.

EMPIRICAL ASSESSMENT OF VALUATION TECHNIQUES Unconditional Analysis The unconditional analysis evaluates the techniques at the average

over all conditions.

Twenty replications are provided by random

assignment of firms to 20 equal-size portfolios in each year without replacement.

The mean number of firms in portfolios (over all years)

was 210, and the mean (over the 20 portfolios) of the (arithmetic) mean portfolio per-share market prices (over years) was $14.29, with a range over the 20 portfolios of $13.79 to $14.65.

The corresponding mean of

the market value of equity was $212.41M (with a range of $192.53M to $230.16M), of carrying value of debt plus preferred stock to the market value of common equity, .90 (with a range of .817 to 1.078), and of estimated beta, 1.13 (with a range of 1.12 to 1.14).

The mean ex ante

CAPM required return on equity was 12.8% (with a range of 12.7% to 12.9%).

Thus the randomization produced portfolios with similar average

characteristics with little variation, including risk attributes. Panel A of Table 1 presents means of portfolio ex post cum-dividend prices, dividends, free cash flows and GAAP cum-dividend earnings (available for common), for selected t+T, all in units of portfolio price at t.

Standard deviations of the means over portfolios

are given in parentheses to give an indication of the similarity of results over the twenty replications. row are calculated as

Pct+T = P t+T +

∑ρ τ =1

Cum-dividend prices in the first T-τ

d t+τ and thus, with the

deflation, the reported values are one plus the stock return. dividends in the table include cash from stock repurchases.

The

Cum-dividend

21 earnings are calculated as

Xt+T + ( ρ -1) ∑ ρ T- τ -1 dt+τ which, when aggregated τ =1

from t to t+T, gives the target in CM (11).

With the deflation, these

give, for each t+T, the cum-dividend earnings yield per dollar of price at t.

All numbers include liquidating amounts for non-survivors (as

described in Appendix B). It is clear that, on average, ex post cum-dividend prices increased more than at the calculated average ex ante rate of 12.8% per year indicated at the bottom of the panel.

This could indicate a

misspecification of this rate but also reflects the bull market of the sample period.

In other words, the data period is not long enough to

average out deviations of realizations from expectations.

Accordingly,

systematic errors that cannot be diversified away by the averaging will be observed for any valid valuation technique.

For the conditional

analysis, valuation errors will be evaluated relative to each other so this is only a concern if portfolios reflect different sensitivities to the systematic departure from expectation. The t+1 figures for dividends, free cash flows, and earnings indicate that the average annual yield of these payoffs was less than the 12.8% rate during the period, but each increased at the average over t+T at a rate greater than 12.8%, consistent with the growth in cum-dividend prices.

However, the increase was less than that of the

ex post prices, consistent with the standard observation that "prices lead" payoffs.

The yields of ex post dividends and free cash flows were

less than that of GAAP earnings.

As free cash flows are returns to

debt, preferred and common equity (whereas earnings are "available to common") it appears that GAAP earnings are closer to the expectation of

22 payoff in the time t price (by which these realizations are initialized) than dividends or free cash flows. Panel B of Table 1 demonstrates this more explicitly.

It gives

mean valuation errors for various valuation techniques for selected horizons.

These valuation errors are per unit of price at t, calculated

as

[

]

Error T (•) = P pt - PTpt (• ) / P pt

(17)

T where P pt (⋅) is the portfolio intrinsic value at t calculated from

ex post realizations to horizon t+T, and Ppt is the observed portfolio price at that date.

Portfolio intrinsic values were calculated

alternatively from means of individual firm's values and by applying the technique to portfolio realizations at each t+T.

The former permits an

examination of firm deviations from means but the mean is sensitive to outliers.

The results here and elsewhere are based on the latter

approach and are similar to the former. The first line in Panel B calculates valuation errors by c

T specifying P pt (⋅) = ρ-T (P pt+T ).

These are errors to forecasting horizon

cum-dividend price at ρT Ppt , that is, by applying the cost of capital to actual price at t.

They are thus the market's forecasting errors, and

we refer to them as price model forecasts.

The negative errors reflect

on-average market inefficiencies at t, misspecification of ρ, or systematic (undiversifiable) ex post deviations from expectation in the period.

Accordingly, they are presented as benchmark errors that arise

for any of these reasons and which one would expect to observe for a

23 perfect valuation technique.

They serve to rescale the calculated

errors for the various techniques.

They may reflect market

inefficiencies (at the portfolio level) at t+T also and these are not anticipated by the valuation techniques. Rows two through five of the panel give valuation errors for the dividend discount model (DDM), the discounted cash flow model (DCFM), the residual income model using GAAP earnings and book values (RIM), and the capitalized GAAP cum-dividend earnings model (CM).

These are

calculated according to equations (1), (8), (9), and (11), respectively, with the target projected to the relevant t+T without a terminal value. The DCF calculation follows the conventional one of specifying FAt as negative and equal to debt plus preferred equity (measured at their carrying values).8

Free cash flow is after income taxes so the tax

benefit of debt is included.

Errors for the DCFM and RIM with terminal

values are given lower in the panel.

These are calculated according to

(14a) with S = 1 and K1 , the annual "growth rate," set to 1.0 and 1.04 for the DCF model (for going concerns) and 1.00 and 1.02 for the RIM model, as indicated.9

Finally, the results for a dividend discount

~ ∑ ρ E(d

) + ρ [(ρ - K )

model calculated with a terminal value as

Pt =

-τ

τ =1

t+τ

-T

1

~ E( d t+T+1)

-1

]

(DDMA)

(18)

are also reported (with K1 = 1.00 and K1 = 1.04). The errors for the dividend discount models are large and positive for short horizons but decline over t+T towards the benchmark errors as more dividends (including liquidating dividends) are "pulled in" to the

24 calculation.10

The errors for the DCF calculation are also positive and

large over all horizons, indeed greater than 150% of actual price. These errors reflect the missing accounting for operations.

With the

terminal value calculations, the errors are still large for all t+T, though declining with higher values of K1 . mean error for t+8 was -0.076.)

(When K1 was set to 1.06 the

In contrast, the errors based on GAAP

accounting in RIM and CM are lower for all horizons and much closer to the benchmark errors, reflecting the accounting for operating assets. Interpreted differently, a DCF calculation with capitalized GAAP operating income as a terminal value performs better than one based on capitalized free cash flow (calculation (15d) versus (16)). The performance rankings are similar with the different calculations of the cost of capital.

Mean absolute deviation of

portfolio errors from these means were also calculated and the rankings over techniques were similar to that for means.

In no case did earnings

methods yield lower bias with higher variation in errors.

B.

Conditional Analysis The results in Table 1 pertain to the market portfolio and the

reported errors are systematic errors.

Valuation also involves

distinguishing firms from the market and we now examine how errors differ over firms (for varying horizons) when the alternative techniques are applied.

The analysis proceeds as before except that firms are

assigned to 20 portfolios each year from a ranking on a conditioning variable that captures the accounting of the various techniques.

25 The use of the accounting models is justified by the difficulty of applying the DDM over finite horizons. firm has no or low payout.

So, first, we assigned firms to portfolios

based on payout to price at time t. request.

This difficulty is acute when a

Detailed results are available upon

Predictably, the DDM and DDMA valuations varied over payout

and this is demonstrative of the problem:

variations in payout (over

finite horizons) that produce different calculations are irrelevant to ex ante values.

Errors for short horizons were typically large.

Those

for the DCF techniques were also large for all horizons, though declining in payout.

In contrast the RIM and CM methods produced

considerably lower errors over all levels of payout. The main focus, however, is on the horizons that the alternative accounting techniques typically require.

That is determined by their

accounting and so we group firms on features of the accounting.

We

identify conditions where a particular technique performs poorly or well and how competing techniques perform under the same conditions.

The

accounting is defined by measurement rules for the stocks and flows so our analysis examines valuations for groups with different measures of the stocks and flows.

26

B.1

Conditioning on the Current Stock Accounting We first group firms on the current stock variables (Bt ) of the

respective techniques.

A special case of the generic accounting model

in (3) (and of the finite horizon model in (14)) arises when the accounting system accounts for Bt such that Pt = Bt (and the other terms in (3) and (14) are zero).

Here the horizon is T=0, all the future is

brought forward into the current book value, and current book value is sufficient for all expected future payoffs (by applying the cost of capital to the book value).

Clearly this "market value accounting" is

an ideal case for practical valuation analysis.

To the extent this is

not satisfied, there is missing value in the current stock and one has to project the future to discover this value, and thus T>0.

The ratio

of the time-t stock to price captures the missing value, so we rank firms on this ratio for DCF accounting and GAAP accounting and examine the implied horizons (to capture the missing value) over deviations from the ideal. Table 2 gives mean errors of the various techniques for 20 portfolios formed from ranking firms on FAt /Pt .

FAt is the DCF stock and

is again measured as (minus) the carrying value of debt plus preferred stock (PS).

Only results for horizons t+1, t+5 and t+8 are reported;

those for intervening horizons are approximated by rough interpolation. The layout of the table is a template for subsequent tables. valuation errors for six models are given as indicated.

Panels of

Results with

alternative calculations of terminal values are available upon request. The table also reports the mean of the ranking variable,

27 (Debtt + PSt )/Pt for each portfolio, the GAAP B/P ratio at t and free * e * e cash flow to equity, FCF t/Pt where FCF t ≡ Ct - It + i t (with i

negative and equal to the after-tax interest on debt plus preferred dividends), and the GAAP E/P ratio at t.

These are ranking variables in

subsequent tables and this table displays their relationship to the ranking variable here. The errors from predicting cum-dividend price by applying ρT to current price (the price model in the first panel) are negative and reflect the systematic unexpected value appreciations documented earlier.

Differences in relative performance is indicated across

portfolios with very high leverage firms performing better than average, demonstrating the effect of (favorable) leverage in good times.

These

errors provide the benchmarks for each portfolio. The ranking variable compares the stock variable in the DCF calculation to price.

Clearly, price cannot be equal to debt plus

preferred stock, but, as price equals the value of operating assets minus the value of debt plus preferred stock, the ranking ratio captures the value of the omitted operating assets in the DCF stock.

Over all

levels of this condition the DCFM errors are positive and large for all horizons and are positively related to the level of omitted operating assets to price. errors. at t.

They are also negatively related to the benchmark

The payoff in free cash flow is too low to justify the price The low FCF after debt service is of course due to high

investment relative to cash from operations, and this is extreme in the case of the high debt firms.

The "terminal value correction" with

K1 = 1.04 reduces these errors but they are still large and the

28 relationship to omitted operating assets remains.11

The results for

portfolio 1 are similar to those for dividend discounting (not reported) as these are pure equity firms where free cash flow equals dividends. The valuation for portfolios 2 to 20 implicitly involves adjusting forecasted dividends for forecasted financing, by CCE, but the ex post errors are larger with this adjustment. As the zero-horizon ideal of Pt = Bt is not possible with DCF accounting, one has to forecast future free cash flows but the results indicate that this calculation does not bring the future forward within horizons less than nine years. operating assets.

GAAP book values include a measure of

Correspondingly, the errors of the RIM calculation

are much closer to the benchmarks.

They are in the order of the

benchmarks but still higher, indicating value payoffs are not entirely captured by the accounting.

The terminal value calculation

(RIM(TV:1.0)) reduces the errors for the lower portfolio numbers, but increases them for the higher ones (as explained with the next table). The CM errors also are lower than DCFM but are typically higher in the extremes.

Mean square and mean absolute deviation of errors are

calculated for each portfolio (overtime) and these are also considerably larger for the DCF calculation than the GAAP ones. While the GAAP calculations are an improvement over DCFM in Table 2, their errors relative to the benchmarks are not perfect.

In

Table 3, firms are ranked on GAAP B/P (the GAAP stock to price) and this gives a spread relative to the ideal of Pt = Bt . identified with portfolio 13 in the table.

This ideal is

The negative correlation

between B/P and the price model errors describes the positive

29 correlation between B/P and subsequent beta-adjusted returns documented in Fama and French (1992), among others.

This could indicate superior

ex post performance for high B/P firms or higher risk, but also may reflect the often-claimed market inefficiency in pricing book values; we just take them as benchmark errors that reflect any of these phenomena.12 The valuation errors for RIM are positively related to the deviation of B/P from unity in portfolio 13.

However, those for high

B/P are close to their benchmark errors for t+5 and t+8.

It is the low

B/P firms for which the errors are relatively high and, as the ex ante error for RIM is given by E(t+T is persistently low up to t+8.

t+T

), these are firms for which the B/P

The RIM (TV:1.0) calculation in part

supplies the missing value for the low B/P firms (and of course more so with a growth rate), but its errors for high B/P are actually higher than those for RIM.

These are portfolios which on average had negative

residual income and capitalizing a negative amount in the terminal value calculation reduces the valuation.

This is of course a legitimate

calculation as firms can have negative residual income (return on equity less than the cost of capital) perpetually and accordingly trade persistently at a discount to book value.13

However, the results

indicate that the horizons for the firms are too short and that the negative residual incomes expected at t will ultimately be higher.14 The errors for CM are also ordered on the benchmarks except they are higher for both low and extremely high B/P firms.

The error of this

model is explained by changes in premiums and it is indeed the extreme B/P that are associated with the biggest changes in premiums (Ou and

30 Penman (1995)).

The errors for DCFM without a terminal value are very

large (and positive) and we don't report them in this or subsequent tables.

It is clear from the DCFM (TV:1.04) results reported that DCF

analysis, even with a growth rate of .04 for the horizon correction, produces no remedy under these conditions.

This is expected given the

positive correlation between (Debt + PS)/P and B/P, because the table also indicates that FCF tends to be negative for low B/P firms. The results for portfolio 13 (where book value approximates price) provide a particular point of reference.

Here one expects cum-dividend

price and book value to grow at the cost of capital and accordingly firms to earn cum-dividend earnings at the cost of capital (zero residual income).

Thus portfolio 13's RIM valuation errors, just like

those for the price model, represent systematic unexpected errors due to ex post rather than ex ante phenomena.

Accordingly its RIM errors

provide an alternative benchmark that reflects the unexpected ex post errors due to unexpected value appreciation.

The errors for t+5 and t+8

are higher than those for the price model and this is consistent with the phenomenon that "price leads earnings":

unanticipated value changes

are incorporated into price before being recognized in earnings and book value.

Errors for other portfolios reflect the phenomenon and thus

should be scaled for it.

31

B.2

Conditioning on the Current Flow Accounting Rather than the current stock being sufficient for valuation, the

current flow, Xt , might be sufficient such that all expected future flows are projected by applying the cost of capital to the current flow. Adding dt to both sides of (11) and substituting

∑ρ τ =1

τ

Xt for the

expected cum-dividend earnings in that expression (to give the projection from current earnings), (11) reduces to

Pt + d t =

ρ X, ( ρ -1) t

(19)

that is, cum-dividend price is the capitalized current flow and the (Pt + dt )/Xt ratio is determined solely by the cost of capital.

Under

this ideal all the future is pulled into the current flow calculation and the horizon is zero.15 In Table 4 firms are ranked on FCFe /P and in Table 5 on the GAAP E/P at t, both with cum-dividend prices in the denominator.

The ranking

maximizes the dispersion from the ideal (for Xt ≡ FCFt to equity and Xt = GAAP earnings to common at t in the two tables).

Portfolios 15 and 16

in Table 4 have FCFe /P closest to the ideal in (19) ((ρ-1)/ρ = .128/1.128 = .113) given the sample's average cost of capital of 12.8%. The errors for DCFM (TV:1.04) are indeed relatively small for these portfolios but increase over portfolios as FCFe /P deviates from this value, and in a direction opposite to those for the price benchmark errors.

They are particularly high for negative FCF firms where the

problem of using DCF analysis is acute.

The errors for RIM, with and

32 without the terminal value, are much lower but, as with those for CM, they are higher for portfolios where the reported B/P are low. In the results based on GAAP E/P rankings in Table 5, the CM(GAAP) model provides a benchmark reference.

Portfolios 14 and 15 have mean

E/P closest to .113 and thus represent the ideal in (19).

By the same

logic that P/B = 1 provided a benchmark in Table 3, the CM errors for these portfolios provide a benchmark that reflects ex post errors adjusted for errors expected given the systematic unpredictable value appreciation.

Indeed the errors for portfolios 14 and 15 for CM are

quite similar to those for RIM in benchmark portfolio 13 in Table 3. The CM errors increase from this benchmark as the spread from the ideal increases, and in a direction consistent with the price model errors.16 However, they are higher for low E/P portfolios. RIM calculations with and without terminal values. again large.

This is so for the The DCFM errors are

33

B.3

Conditioning on Accruals The difference between free cash flows and GAAP earnings is the

GAAP accruals for operating activities.

Referring to Section I.D., GAAP

* earnings, Xt = Ct + i t + oat (where oat are operating accruals).

As

* equity FCFt = Ct - It + i t, the difference between the two flow measures

is It + oat .

The FCF calculation treats investment as an immediate

diminution of value by signing it negative.

This is commonly claimed as

the reason DCF analysis requires long horizons:

as investment enters

negatively into the calculation, a long horizon is required to capture the subsequent cash (in flows) from the investment. Table 4.

This is apparent in

For extreme FCF portfolios, where investment or disinvestment

is large relative to cash flow from operations, the DCF valuation errors are particularly high.

Accrual accounting (in general) treats

investment as an operating asset that does not immediately affect earnings17 and, in addition, brings other future cash flows forward in time through operating accruals (oa) like receivables and pension liabilities.

The results above indicate that this accounting reduces

the error in DCF valuation.

However, accruals are by fiat and may

themselves introduce error. Table 6 ranks firms on GAAP E/P minus FCFe /P.

This difference is

equal to (I + oa)/P which is also the change in operating assets to price.

The greater the absolute difference between FCF and GAAP

earnings, the worse the DCF calculations perform.

In contrast the

calculations made from GAAP attributes produce errors considerably closer to the benchmarks.

Significantly, the accrual accounting

34 produces the largest correction to DCF analysis when free cash flows are extreme and when the difference between earnings and free cash flows is the highest.

The treatment of investment and additional accruals in

GAAP accounting serve to correct the FCF calculation to facilitate finite horizon analysis.

The results suggest that rather than adjusting

earnings forecasts to get back to cash flows, one is better served (for valuation purposes) to preserve the accrual accounting.

B.4

Conditioning on GAAP B/P and E/P The evidence indicates that GAAP accounting facilitates practical

(finite horizon) valuation better than DCF calculations.

However, the

results also indicate conditions where GAAP models do not perform well relative to benchmarks.

These are cases of high and (particularly) low

B/P (Table 3) and low E/P (Table 5).

These conditions are associated

with the central portfolios in Table 6 where again the RIM errors are the highest.

Given that these findings are not due to market

inefficiencies, misspecification of discount rates or GAAP violations of clean surplus accounting, then the zero error conditions for (3), (11) and (14) are not satisfied in these circumstances for the horizons investigated. The results for B/P and E/P involve conditions where the accounting produces extremes relative to the ideals.

The results for

high B/P were explained by ex post negative residual income.

As the

other conditions involve low book values and earnings to price, one suspects that conservatism in the accounting might be affecting the

35 calculations.

Conservative accounting that writes down book values is

reputably present in the prescription and practice of GAAP accounting. Median B/P ratios are less than unity in the sample, reinforcing this impression.

The low B/P are cases where the conservatism is likely to

be extreme, and indeed the expensing of R&D expenditures under SFAS No. 2, for example, is associated with low B/P.

However, conservative

accounting for book values of assets is not sufficient to violate the horizon conditions in (14) from RIM (TV:1.0) as conservatism, consistently applied, will in this case produce the constant premiums on which the calculation is predicated.

The no-change-in-premium-condition

implies expected cum-dividend earnings equal to expected cum-dividend price change (return) at the horizon, that is, the conservatism does not affect earnings relative to price. describes this:

Standard textbook accounting

rapid amortization does not affect earnings if there is

no change in the asset base because depreciation expense is the same whether one expenses an investment immediately or capitalizes and amortizes it.

However, if amortizable assets are expected to be

changing at the horizon, expected earnings will be affected by conservative accounting (downwards for growing assets), and the constant premium condition for K1 = 1.0 will not be satisfied.

This has a formal

representation in Feltham and Ohlson (1995). If low earnings relative to price at t and low book value relative to price at t are indicative of low values of the two accounting numbers to price at the horizon,18 then one expects these effects to be identified by current B/P and E/P, as indicated.

As the effect is

induced by the accounting in both measures, the conditioning

36 circumstance involves joint values of the two ratios.

Accordingly,

Table 7 displays mean valuation errors for various joint values of B/P and E/P.

In each year firms are ranked on B/P and those with

.95 ≤ B/P < 1.05 assigned to portfolio 12.

Then firms with B/P < .95

are ranked on E/P and assigned to portfolios 1-11 from this ranking, and those with B/P ≥ 1.05 are also ranked on E/P and assigned to portfolios 13-20.

Portfolio 12 has a mean B/P of .996, close to unity by

construction.

However it also has a mean E/P of .108 (and a similar

book return on equity by implication), closest to the assumed cost of capital.

Thus this portfolio describes results for both a normal book

value and a (close to) normal P/E ratio in (19) and the errors for the RIM and CM calculations (highlighted in the table) are again the benchmark errors given the systematic ex post price errors.

For low B/P

(portfolios 1-11), the errors for RIM and CM are increasing in decreasing E/P:

the mispricing of these models is identified with both

low B/P and low E/P.

The GAAP models do not perform well in conditions

that are associated with conservatism in the accounting for book values and its spillover to earnings when assets are growing. This deficiency is also apparent for RIM (TV:1.0) where capitalized terminal residual income is too low for these conditions. The K1 captures the measurement error in earnings induced by the accounting and expected growth in residual income at the horizon is determined by it.

Accordingly, the specification of K1 > 1.0 provides

an accommodation.

Table 7 reports results for RIM (TV:1.04) and its

errors for low B/P and E/P portfolios are considerably lower than those of RIM (TV:1.0) for the longer horizons.

However, even with this

37 adjustment, the errors are higher than the benchmarks.

If one considers

a perpetual growth in residual income higher than 4% to be unreasonable, then longer forecasting horizons are required for these firms.

In any

case, it is clear that DCFM (TV:1.04) provides no remedy in these conditions.

V.

CONCLUSION The paper documents that equity valuation methods based on

forecasting GAAP (accrual) earnings and book values have practical advantages over dividend discounting and discounted cash flow analysis. GAAP accounting has the feature of bringing the future forward in time in accruals and, by an accounting for operating assets, excluding investment expenditures as a charge against cash flow from operations in the accounting for the payoff.

This facilitates valuing firms from

forecasts of payoffs over relatively short horizons. The analysis of valuation errors of the relevant techniques over different conditions provides a practical guide to when a particular technique will work well (or otherwise).

The analysis is couched within

a framework that reveals the accounting at work in the contrasting techniques, so errors can be identified with the missing accounting in a particular technique.

Thus the results indicate that GAAP accounting

supplies some of the missing accounting (for operations) in DCF analysis, but they also indicate conditions (associated with high P/E and high P/B firms) when the GAAP accounting is unsatisfactory.

GAAP

accounting is of course only one form of accrual accounting and one

38 might investigate other rules that provide a remedy.

Indeed the

analysis suggests a utilitarian criterion for normative accounting principles:

they should facilitate finite-horizon valuation.

In this respect the analysis that compares pure cash flow accounting with GAAP accrual accounting is quite narrow and further research might compare alternative accrual accounting systems. might promote better techniques and better accounting.

This

Also, the paper

takes a macro view by looking at average results in the cross section. One might continue the analysis in micro settings (for industries or firms) and, as the accounting for operations is an important feature, these might be identified by characteristics of firms' operations. DCF techniques are the most common in practice and in teaching in business schools.

The typical valuation book "backs out" accruals from

financial statements to get to the cash flows.

In Copeland, Koller and

Murrin (1990), for example, "cash is king" and the accounting is suspect.

The results here modify that view.

However, the paper shows

that DCF techniques that involve (accrual) operating income in the terminal value calculation are equivalent to residual income accrual accounting techniques.

In effect, this terminal value calculation

corrects the errors from forecasting free cash flow to the horizon to get back the accruals.

One questions the efficiency of going through

this exercise (of taking out accruals and then adding them back in) when forecasting the accrual numbers produces the same result, and indeed whether this can really be called a DCF technique.

39

H-THS.17-13N

40

Appendix A Calculation of Target Attributes This appendix provides details of calculations in the implementation of the alternative valuation approaches.

Dividends Dividends are defined as common dividends per share ex date (COMPUSTAT item 26) adjusted for stock splits and stock dividends over time.

For the dividend discount model, per-share cash distributions

from stock repurchases were added to these dividends at each t+τ to capture all cash payments to shareholders.

Since information about

stock repurchases is not available in our data bases we followed Ackert and Smith (1993) and Shoven (1986) to discover them.

We searched the

CRSP monthly returns file for information on shares outstanding and each decrease in shares outstanding (adjusted for stock splits and reverse splits) was treated as a stock repurchase.

The amount of cash

distributed was determined by multiplying the decrease in shares by the price at the end of the month preceding the decrease.

This amount was

divided by the number of shares outstanding before the decrease to arrive at a per-share cash distribution. For the period 1973 to 1992 we detected 7,659 share decreases (6,117 for NYSE and AMEX firms and 1,542 for NASDAQ firms).

This number

seems plausible for our sample period given that Comment and Jarrell (1991) report 1,303 stock repurchases for the 4-year period 1985-1988.

41 The mean per-share distribution was $0.99 with a standard deviation of 3.82 over the 20-year sample period.

Free Cash Flows Two free cash flow calculations were made.

Results are similar

for the two calculations, but those reported are based on the second.

a) Calculations Based on GAAP Free Cash Flow Accounting regulation for the reporting of cash flow information first appeared in 1971 when APB Opinion No. 19 mandated the preparation of the "Statement of Changes in Financial Position."

Under this

regulation firms reported a working capital statement (COMPUSTAT format code = 1.000), a cash by source and use of funds statement (COMPUSTAT format code = 2.000), or a cash statement by activity (COMPUSTAT format code = 3).

In 1987 the FASB issued Standard No. 95, "Statement of Cash

Flows" mandating the reporting of cash receipts, cash payments, and net change in cash resulting from operating, investing, and financing activities during a period (COMPUSTAT format code = 7.000). For firms with format code = 7.000 in a given year, cash from operations (Ct+ τ ) was calculated as Ct+ τ = Operating Activities-Net Cash Flow (item 308) + Interest Paid-Net (item 315). When Interest Paid-Net was not available, then Interest Expense (item 15), if available, was substituted for Interest Paid-Net. Cash Investment (It+ τ ) was calculated as It+ τ = Investing Activities-Net Cash Flow (item 311)

42 + Capitalized Interest (item 147). Investments in financial assets are included in investing activities under GAAP.

We did not exclude them as, for the DCF calculation, we

identified only debt and preferred stock as financial items. For format codes 1, 2 and 3, Ct+ τ - It+ τ = the change in Cash and Cash Equivalents (item 274) - Sale of Common and Preferred Stock (item 108) - Long-Term Debt-Issuance (item 111) + Long-Term Debt-Reduction (item 114) - Change in Current Debt (301) + Purchase of Common and Preferred Stock (item 115) + Cash Dividends (item 127) + Interest Expense (item 15) + Capitalized Interest (item 147). When data were not available for this calculation, Ct+ τ = Funds from Operations (item 110) + Interest Expense (item 15) - Working Capital Changes-Other (item 236). It+ τ was calculated by one of the following: 1 = Increase in Investments (item 113) I t+τ + Capital Expenditures (item 128) + Acquisitions (item 129) - Sale of Investments (item 109) - Capitalized Interest (item 147). or 2

I t+τ

= Change in Property, Plant and Equipment-Total (Net) (change in item 8) + Depreciation and Amortization (item 14) + Change in Investments and Advances-Other (change in item 32) + Change in Intangibles (item 33) - Capitalized Interest (item 147).

When a format code was not available, Ct+ τ = Income Before Extraordinary Items (item 18) + Extraordinary Items and Discontinued Operations (item 48) + Depreciation and Amortization (item 14) + Interest Expense (item 15) + Change in Deferred Taxes (item 74) - Change in Working Capital (item 179)

43 The change in working capital was modified for the change in Debt in Current Liabilities (item 34) when available.

Cash Investment was

2 calculated by I t+τ above.

In all calculations, items not available were set to zero.

b)

Calculation Based on Articulation of Balance Sheet and Income Statement The above calculations are complicated and fraught with

difficulties due to nonavailability of some line items. I.D, ∆OAt+ τ = It+ τ + oat+ τ and oat+ τ = OIt+ τ - Ct+ τ .

From Section

Thus Ct+ τ - It+ τ = OIt+ τ -

∆OAt+ τ and free cash flow is calculated by identifying operating income in the income statement and net operating assets in consecutive balance sheets (Feltham and Ohlson (1995)). Thus we calculated Ct+ τ - It+ τ = Income Before Extraordinary Items (item 18) + Interest Expense (item 15) - Change in Total Assets (change in item 6) + Change in Total Liabilities (change in item 181) - Change in Total Long-Term Debt (change in item 9) - Change in Debt in Current Liabilities (change in item 34) The exclusion of extraordinary items excludes gains and losses on debt repurchases (financing activities) but also may exclude some operating activities. in portfolios.

This is a problem only if these are not mean zero

By this calculation (as in the first), financial assets

(FAt ) are identified as (minus) the sum of debt and preferred stock at time t measured at their carrying values, as designed.

The book value

of debt is the book value of long-term debt (COMPUSTAT item 9) plus the debt in current liabilities (COMPUSTAT item 34). COMPUSTAT item 130 is used.

For preferred stock

44 These calculations of free cash flow are after tax but include tax benefits of interest on debt.

Accordingly the "value of the tax shield"

is included in the present value calculation rather than being calculated separately (as in the "compressed adjusted present value technique" employed by Kaplan and Ruback (1995)). The calculation of free cash flow and financial assets is on a total dollar basis and the total dollar intrinsic price at t was placed on a per-share basis by dividing by shares outstanding at t (item 25).

GAAP Accounting GAAP earnings and book value were calculated on a per-share basis. Earnings at time t+τ were identified as primary earnings per share (COMPUSTAT item 53), adjusted for stock splits and stock dividends over time.

Book value per share is influenced by share issues so its value

at each t+τ (for determining residual income) was calculated as book value per share at t plus accumulated earnings net of cash dividends per share from t to t+τ (split-adjusted).

In calculating the payoff for the

CM in (11), dt+ τ is specified as common dividends per share ex date, adjusted for stock splits and stock dividends over time.

That is, dt+ τ

does not include cash distributions from stock repurchases as these are reflected in the per-share calculation of earnings per share.

45

Appendix B The Treatment of Terminations

Firms may delist for various reasons including mergers, liquidations, acquisitions, insufficient reporting, etc. and the analysis incorporates corrections for ex post terminations in each of the valuation models.

The information needed for such corrections is

obtained from the CRSP monthly returns files and the COMPUSTAT files. The CRSP files provide delisting codes indicating the reason for delisting.

We detected 3,355 delistings for our sample in the period

1973-1992 (1,792 delistings for NYSE and AMEX firms and 1,563 for NASDAQ firms).

Out of this total, 1,851 delistings were due to mergers, 261

due to acquisitions, 124 due to liquidations and the remaining 1,119 due to other reasons.

In addition, the CRSP files provide per-share amounts

for any cash or non-cash distributions at the time of termination.

We

detected 1,736 cash and 1,013 non-cash terminal distributions for our sample (some firms had both cash and non-cash distributions).

The mean

terminal distribution was $19.85 per-share with a standard deviation of $20.62.

A total of 402 NYSE and AMEX firms and 797 NASDAQ firms did not

have any cash or non-cash terminal distributions.

In this case the

terminal (non-cash) distribution was assumed to be equal to the last price of the delisting firm in the CRSP files.

The mean terminal price

for these firms was $3.55 with a standard deviation of $8.50. The corrections for terminations incorporated in each valuation model are given below.

46

Dividend Discount Model (DDM) In this model the dividend in the termination year was measured as the sum of dividends per share, cash distribution per share from stock repurchases and cash and/or non-cash terminal distribution per share, all adjusted for stock splits and stock dividends.

Tracking subsequent

dividends on non-cash distributions is difficult so the assumption is that securities received are liquidated.

Discounted Cash Flow Model (DCFM) For terminating firms, intrinsic values were calculated as the present value of dividends (including terminal dividends) as this equals the present value of cash flows.

Earnings Models (RIM and CM) To estimate earnings in the terminal year we followed Bernard (1994) and estimated terminal gains or losses as the difference between the last market value on CRSP and the last book value on COMPUSTAT.

The

mean per share terminal gain for the 3,355 delistings in our sample was $6.46 with a standard deviation of 16.68.

Earnings Xit+ τ in the

termination year was calculated as the sum of reported earnings per share and the terminal gain or loss per share.

47 FOOTNOTES

An exception is Kaplan and Ruback (1995). This contrasts with Kaplan and Ruback (1995) and Abarbanell and Bernard (1995) where prices are compared to values calculated from forecasts of cash flows or earnings. That approach is limited by the availability of dividend and cash flow forecasts and of earnings forecasts for longer horizons. Further, it assumes that the analyst forecasts identified are unbiased. On this point see Frankel and Lee (1995).

Rubinstein (1976) derives the model under no-arbitrage conditions. In that derivation the discount for risk is in the numerator which is then discounted to present value at the riskfree rate. The common textbook form is stated here as this is usually how the model is applied in practice. DCF analysis is sometimes applied with adjustments (distinguishing "discretionary" capital investments, for example). These amount to additional accounting of which GAAP accrual accounting for operational activities is one form, and the aim here is to compare this accounting with the strict cash accounting case. Firms with different fiscal year ends in the same calendar year were assigned to portfolios together and portfolio prices were based on firm prices of fiscal year end. The Fama and French estimates are based on risk premiums estimated on data after our portfolio formation dates. For CAPM rates, we also used an 8% equity premium which was the historical rate at the end of the sample period. (This of course produced lower errors as we define them.)

Precisely, those used in Kaplan and Ruback (1995) with unlevered betas calculated from estimated equity betas and debt and preferred stock betas assumed to be 0.25. The income tax rate was set at the prevailing top federal corporate rate in the relevant year plus 4% for state and other taxes.

Financial assets are not netted out as one has difficulty distinguishing them from operating cash in cash and cash equivalents. Accordingly interest income (but not expense) is included in free cash flows. The analysis was also repeated with growth rates of 1.02 and 1.06, with similar results. A rate of 1.04 or 1.06 applied to RIM, though reducing errors towards the benchmark might be considered excessive: they imply a relatively rapid perpetual growth in book return on equity.

For companies that terminated, the liquidating dividend was calculated as the price at liquidation (see Appendix B). To the extent that distributions in liquidation were stock rather than cash, calculated values for the DDM are overstated. The occasional large negative value in this panel arises from capitalizing relatively high free cash flow in the terminal value calculation with the small capitalization rate that K1 = 1.04 can produce.

The results in Table 3 were similar when the cost of capital was based on the Fama and French

48

(1994) risk model which includes B/P as a risk factor. Ou and Penman (1995) document persistent discounts. In the DCF terminal value calculations we capitalized only positive free cash flows as a perpetual negative free cash flow is not realistic. The ex post negative residual earnings could also be due to our specification of the cost of capital on which the residual earnings calculation is based. The negative amounts were particularly associated with periods of high interest rates.

This is what Fischer Black had in mind when he advocated calculating earnings as a sufficient number (with a multiplier) for value. See Black (1980). Note, however, that the price model errors might reflect the so-called "P/E effect" pricing inefficiency. An exception is the accounting for investment in research & development under SFAS No. 2. Beaver and Morse (1978) document that differences in P/E ratios from the median persist over time and Ou and Penman (1992) provide similar documentation for premiums.

49 REFERENCES Ackert, L. and Smith, B. "Stock Price Volatility, Ordinary Dividends, and Other Cash Flows to Shareholders." Journal of Finance (September 1993), 1147-1160. Abarbanell, J. and Bernard, V. "Is the Stock Market Myopic?" Paper, University of Michigan, 1995.

Working

Beaver, W. and Morse, D. "What Determines Price-Earnings Ratios?" Financial Analysts' Journal (July/August 1978), 65-76. Bernard, V. "Accounting-Based Valuation Methods, Determinants of Market-to-Book Ratios, and Implications for Financial Statement Analysis." Working Paper, University of Michigan, 1994. Black, F. "The Magic in Earnings: Economic vs. Accounting Earnings." Financial Analysts Journal (November/December 1980), 19-24. Comment, R. and Jarrell, G. "The Relative Signalling Power of DutchAuction and Fix-Price Self-Tender Offers and Open-Market Share Repurchases." Journal of Finance (September 1991), 1243-1271. Copeland, T., Koller, T., and Murrin, J. Valuation: Measuring and Managing the Value of Companies. New York: John Wiley & Sons, 1990. Cornell, B. Corporate Valuation: Tools for Effective Appraisal and Decision Making. Homewood, IL: Business One Irwin, 1993. Easton, P., Harris, T., and Ohlson, J. "Aggregate Accounting Earnings Can Explain Most of Security Returns." Journal of Accounting and Economics (June/September, 1992), 119-142. Edwards, E. and Bell, P. The Theory and Measurement of Business Income. Berkeley, CA: The University of California Press, 1961. Fama, E. and French, K. "The Cross-Section of Expected Stock Returns." Journal of Finance (June 1992), 427-465. Fama, E. and French, K. "Industry Costs of Equity." University of Chicago, 1994.

Working Paper,

Feltham, G. and Ohlson, J. "Valuation and Clean Surplus Accounting for Operating and Financial Activities." Contemporary Accounting Research (Spring 1995), 689-731. Frankel, R. and Lee, C. "Accounting Valuation, Market Expectation, and the Book-to-Market Effect." Working Paper, University of Michigan, 1995.

50 Hackel, K. and Livnat, J. Cash Flow and Security Analysis. IL: Business One Irwin, 1992.

Homewood,

Ibbotson, R. and Sinquefield, R. Stocks, Bonds, Bills and Inflation, 1926-1982. Charlottesville, VA: Financial Analysts Research Foundation, 1983. Kaplan, S. and Ruback, R. "The Valuation of Cash Flow Forecasts: Empirical Analysis." Journal of Finance (September 1995), 1059-1093.

An

Miller M. and Modigliani, F. "Dividend Policy, Growth and the Valuation of Shares." Journal of Business (October 1961), 411-433. Modigliani, F. and Miller, M. "The Cost of Capital, Corporation Finance, and the Theory of Investment." American Economic Review (March 1958), 261-297. Ohlson, J. "Earnings, Book Values, and Dividends in Equity Valuation." Contemporary Accounting Research (Spring 1995), 661-687. Ou, S. and Penman, S. "Financial Statement Analysis and the Evaluation of Market-to-Book Ratios." Working Paper, University of California, Berkeley, 1995. Peasnell, K. "Some Formal Connections Between Economic Values and Yields and Accounting Numbers." Journal of Business Finance and Accounting (October 1982), 361-381. Penman, S. "A Synthesis of Equity Valuation Techniques and the Terminal Value Calculation for the Dividend Discount Model." Working Paper, University of California, Berkeley, 1996. Penman, S. "The Articulation of Price-Earnings Ratios and Market-toBook Ratios and the Evaluation of Growth." Working Paper, University of California, Berkeley, 1995. Forthcoming, Journal of Accounting Research. Penman, S. and Sougiannis, T. "The Dividend Displacement Property and the Substitution of Anticipated Earnings for Dividends in Equity Valuation." Working Paper, University of California at Berkeley and University of Illinois, 1994. Preinreich, G. "Annual Survey of Economic Theory: The Theory of Depreciation." Econometrica (July 1938), 219-241. Rappaport, A. Creating Shareholder Value: The New Standard for Business Performance. New York: The Free Press, 1986. Rubinstein, M. "The Valuation of Uncertain Income Streams and the Pricing of Options." Bell Journal of Economics (Autumn 1976), 407-425.

51

Shoven, J. "The Tax Consequences of Share Repurchases and Other Non-Dividend Cash Payments to Equity Owners." In Lawrence Summers, ed.: Tax Policy and the Economy, Vol. 1 (MIT Press, Cambridge, Mass.), (1986), 29-54. Stewart, G. The Quest for Value. Inc., 1991.

New York:

Harper-Collins Publishers

Table 1 Mean Portfolio Values of Realized Valuation Attributes (Panel A) and Ex post Valuation Errors for Valuation Techniques (Panel B), for Selected Horizons (Standard Deviation of Portfolio Means in Parentheses) 1973-1990 t+1

t+2

Horizon (t+T) t+4 t+6

t+8

Cum-div. price

1.155 (.006)

1.371 (.018)

1.934 (.048)

2.769 (.081)

3.814 (.112)

5.298 (.228)

Dividends

.088 (.007)

.100 (.008)

.140 (.013)

.181 (.018)

.235 (.036)

.285 (.053)

Free cash flows

.076 (.030)

.108 (.038)

.104 (.051)

.147 (.059)

.206 (.109)

.233 (.105)

Cum-div. GAAP earnings

.100 (.005)

.115 (.006)

.174 (.010)

.236 (.013)

.308 (.026)

.422 (.037)

(1.128)T

1.128

1.272

1.619

2.060

2.621

3.335

Panel A:

t+10

Future Attributes

Table 1 (continued) t+1

t+2

Horizon (t+T) t+4 t+6

t+8

Price model

-.031 (.006)

-.085 (.014)

-.177 (.027)

-.294 (.035)

-.381 (.046)

-.538 (.073)

DDM

.923 (.006)

.845 (.008)

.663 (.016)

.478 (.021)

.283 (.036)

.069 (.045)

DCFM

1.937 (.057)

1.868 (.058)

1.762 (.066)

1.670 (.078)

1.552 (.086)

1.450 (.099)

RIM

.175 (.013)

.176 (.013)

.103 (.019)

.038 (.021)

-.028 (.027)

-.120 (.039)

CM

.199 (.035)

.189 (.034)

.074 (.033)

.022 (.029)

-.031 (.035)

-.113 (.047)

DCFM (TV: 1.0)

1.254 (.184)

1.188 (.155)

1.112 (.142)

.946 (.251)

.782 (.222)

.827 (.353)

DCFM (TV: 1.04)

.918 (.269)

.853 (.224)

.765 (.199)

.558 (.424)

.378 (.342)

.506 (.560)

RIM (TV: 1.0)

.206 (.045)

.192 (.039)

.083 (.061)

.037 (.073)

.008 (.073)

-.164 (.092)

RIM (TV: 1.02)

.058 (.054)

.049 (.046)

-.061 (.073)

-.099 (.086)

-.117 (.087)

-.307 (.108)

DDMA (TV: 1.0)

.574 (.029)

.504 (.039)

.314 (.042)

.132 (.053)

-.061 (.050)

-.295 (.055)

DDMA (TV: 1.04)

.424 (.043)

.356 (.059)

.167 (.058)

-.010 (.070)

-.203 (.064)

-.452 (.073)

Panel B:

t+10

Valuation Errors

NOTES: Means are mean over years of means for 20 portfolios to which firms were randomly assigned in each year, 1973-90. Standard deviations are means of yearly standard deviations of portfolio values. Valuation error is actual portfolio price at t minus model price, deflated by actual price at t. Price model valuation errors are calculated by setting model price equal to the present value of actual ex-post cum-dividend price at each horizon, t+T. DDM refers to the dividend discount model in equation (1) of the text, DCFM to the discounted cash flow model in equation (8), RIM to the residual income model in equation (9) with GAAP earnings and book values, and CM to the capitalized GAAP earnings model in equation (11). TV indicates a terminal value was calculated for going concerns according to (14a) with the assumed

subsequent growth rate in the terminal payoff indicated within the parentheses. DDMA is the dividend discount model with a terminal value calculated according to equation (18). All calculations include terminal distributions to equity holders for nonsurviving firms.

Table 2 Mean Ex post Valuation Errors of Valuation Techniques for Selected Horizons, for Portfolios Formed from a Ranking on Debt Plus Preferred Stock to Price Portfolio

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Portfolio Attributes at t Mean (Debt+PS)/P

Mean GAAP B/P

.000 .007 .034 .072 .116 .165 .221 .286 .359 .443 .538 .654 .791 .964 1.176 1.442 1.789 2.302 3.344 10.962

.640 .550 .471 .546 .570 .615 .669 .734 .772 .816 .896 .941 .985 1.047 1.082 1.115 1.162 1.270 1.409 1.432

t+1

Horizon t+5

t+8

t+1

Price Model -0.024 -0.002 -0.002 -0.013 -0.034 -0.034 -0.030 -0.030 -0.034 -0.044 -0.037 -0.046 -0.036 -0.047 -0.043 -0.032 -0.040 -0.042 -0.052 -0.080

-0.235 -0.025 -0.044 -0.127 -0.163 -0.173 -0.170 -0.183 -0.244 -0.283 -0.268 -0.294 -0.351 -0.311 -0.235 -0.265 -0.349 -0.414 -0.604 -0.861

-0.388 -0.090 -0.116 -0.205 -0.259 -0.302 -0.331 -0.358 -0.385 -0.404 -0.399 -0.493 -0.447 -0.511 -0.313 -0.485 -0.497 -0.678 -1.060 -1.336

Horizon t+5

t+8

DCFM 0.990 1.005 1.060 1.107 1.197 1.278 1.356 1.470 1.545 1.713 1.735 1.957 2.449 2.426 2.586 3.036 3.413 4.002 5.380 9.204

0.947 0.909 1.009 1.037 1.128 1.167 1.293 1.338 1.409 1.530 1.557 1.809 2.227 2.047 2.313 2.703 3.147 3.622 4.448 7.761

Horizon t+5

t+1

t+8

DCFM (TV: 1.04) 0.877 0.789 0.920 0.956 1.005 1.055 1.140 1.205 1.205 1.331 1.281 1.682 1.841 1.922 2.251 2.758 3.219 3.676 4.214 8.307

0.810 0.714 0.927 0.763 1.018 0.767 1.148 0.915 0.768 1.095 1.036 1.205 1.085 0.192 1.455 1.000 1.388 1.058 0.110 -3.251

0.891 0.351 0.575 0.541 0.421 0.450 0.532 0.131 0.570 0.694 0.596 0.874 -0.766 1.367 0.967 1.551 1.717 1.787 0.531 3.542

0.532 0.272 0.463 0.718 0.041 0.217 0.498 0.177 -0.163 0.444 0.085 0.399 -1.711 0.639 0.974 1.577 1.412 1.490 2.521 6.497

Table 2 (continued) Portfolio

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Portfolio Attributes at t Mean FCFe /P

Mean GAAP E/P

.096 .037 .002 .002 .008 -.002 .008 -.002 -.004 .001 -.006 -.003 .011 .016 -.012 .021 -.032 -.012 .042 .539

.075 .071 .070 .078 .082 .085 .090 .096 .096 .096 .103 .099 .097 .095 .096 .087 .085 .068 .014 -.281

t+1

Horizon t+5

t+8

RIM 0.322 0.415 0.515 0.438 0.416 0.356 0.301 0.232 0.182 0.128 0.066 0.016 -0.018 -0.070 -0.100 -0.119 -0.166 -0.247 -0.356 -0.248

0.178 0.343 0.375 0.292 0.273 0.219 0.169 0.114 0.042 0.002 -0.072 -0.074 -0.116 -0.141 -0.148 -0.143 -0.178 -0.250 -0.302 -0.330

Horizon t+5

t+1

t+8

t+1

RIM (TV: 1.0) -0.026 0.272 0.263 0.147 0.162 0.090 0.023 0.025 -0.054 -0.090 -0.173 -0.193 -0.181 -0.249 -0.184 -0.206 -0.276 -0.330 -0.433 -0.393

0.289 0.311 0.353 0.233 0.222 0.196 0.149 0.176 0.089 0.110 0.000 0.103 0.117 0.080 0.152 0.237 0.269 0.303 0.535 1.132

0.009 0.223 0.169 0.134 0.158 0.049 0.058 0.116 0.046 -0.058 -0.108 0.032 -0.069 -0.038 0.006 0.008 -0.129 -0.010 -0.089 -0.049

-0.208 0.150 0.170 0.131 -0.053 0.007 0.029 0.031 -0.046 -0.035 -0.127 0.032 0.041 -0.027 0.039 -0.098 -0.085 -0.038 -0.037 0.322

Horizon t+5

t+8

CM -0.101 0.181 0.286 0.241 0.259 0.164 0.149 0.132 0.018 0.069 0.041 0.071 0.091 0.144 0.144 0.262 0.183 0.388 0.785 2.330

0.000 0.245 0.234 0.154 0.156 0.098 0.064 0.050 -0.025 -0.033 -0.101 -0.049 -0.096 -0.062 -0.033 0.003 -0.029 -0.029 0.102 0.202

-0.181 0.210 0.191 0.067 0.095 0.014 -0.044 0.008 -0.059 -0.071 -0.139 -0.159 -0.112 -0.171 -0.039 -0.083 -0.184 -0.204 -0.255 -0.098

Notes: PS is the carrying value of preferred stock and FCFe is free cash flow to common equity. The GAAP E/P ratio is calculated as Xt /(Pt +dt ) where Xt is GAAP earnings available for common in the portfolio formation year, t. Pt is the common stock price at the end of year t and dt is the annual dividend for year t. GAAP B/P is reported book value of common equity to price at t. See notes to Table 1 for descriptions of valuation techniques and the calculation of the means.

Table 3 Mean Ex post Valuation Errors of Valuation Techniques for Selected Horizons, for Portfolios Formed from a Ranking on GAAP Book/Price Ratios Portfolio

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Portfolio Attributes at t Mean GAAP B/P

Mean (Debt+PS)/P

-.033 .240 .315 .392 .459 .527 .594 .672 .734 .807 .873 .946 1.021 1.114 1.229 1.338 1.530 1.744 2.150 3.302

.818 .279 .293 .361 .471 .540 .634 .780 .772 .906 .945 1.027 1.341 1.475 1.731 1.839 2.241 2.973 2.925 4.290

t+1

Horizon t+5

t+8

Price Model 0.091 0.038 -0.011 -0.015 -0.025 -0.016 -0.013 -0.037 -0.018 -0.028 -0.046 -0.036 -0.056 -0.070 -0.075 -0.069 -0.067 -0.086 -0.111 -0.143

0.002 -0.008 -0.080 -0.125 -0.146 -0.150 -0.164 -0.211 -0.151 -0.200 -0.221 -0.270 -0.326 -0.375 -0.438 -0.374 -0.488 -0.459 -0.518 -0.933

-0.098 -0.059 -0.090 -0.192 -0.218 -0.257 -0.282 -0.431 -0.242 -0.336 -0.391 -0.503 -0.588 -0.605 -0.677 -0.648 -0.701 -0.716 -0.803 -1.282

Horizon t+5

t+1

t+8

t+1

DCFM (TV: 1.04) 1.231 1.077 1.197 1.191 1.199 1.274 1.040 1.339 1.181 1.095 1.307 0.488 0.815 0.653 0.162 0.281 -0.584 -0.898 -0.545 -2.108

1.272 0.753 0.598 0.910 0.900 0.805 0.848 0.715 0.552 -0.024 0.674 0.839 1.171 0.687 0.891 0.612 0.765 0.539 -0.153 1.189

1.372 0.663 0.700 0.631 0.373 0.731 0.531 -0.544 0.203 0.934 1.152 0.574 0.012 0.368 -0.146 1.865 -0.101 1.439 1.755 -3.167

Horizon t+5

t+8

RIM 1.149 0.770 0.657 0.569 0.497 0.424 0.353 0.284 0.212 0.146 0.073 -0.004 -0.084 -0.171 -0.270 -0.375 -0.526 -0.720 -1.035 -1.910

0.789 0.588 0.465 0.374 0.321 0.255 0.180 0.111 0.068 0.010 -0.040 -0.121 -0.198 -0.242 -0.318 -0.353 -0.453 -0.550 -0.698 -1.073

0.597 0.490 0.318 0.221 0.169 0.110 0.073 -0.037 -0.050 -0.125 -0.161 -0.244 -0.316 -0.327 -0.402 -0.416 -0.502 -0.570 -0.686 -0.817

Table 3 (continued) Portfolio Portfolio Mean FCFe /P 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Horizon Attributes at tt+1

Mean GAAP E/P .485 -.026 -.045 -.052 -.056 -.062 -.021 -.045 -.040 -.030 -.036 -.020 -.024 -.012 .009 .062 .028 .079 .099 .428

t+5

t+8

RIM (TV: 1.00) -.083 .035 .055 .065 .071 .081 .088 .094 .096 .099 .107 .107 .112 .113 .113 .103 .102 .090 .061 -.014

0.500 0.498 0.362 0.266 0.245 0.223 0.153 0.094 0.058 0.068 0.041 0.019 -0.053 0.048 0.000 0.131 0.265 0.324 0.671 1.573

Notes: See notes to Tables 1 and 2.

0.457 0.397 0.232 0.150 0.130 0.054 0.038 -0.061 -0.038 -0.030 -0.066 -0.095 -0.201 -0.040 -0.102 -0.032 -0.136 0.021 -0.030 0.173

Horizon t+1

t+5

t+8

0.415 0.393 0.245 0.163 0.142 0.105 0.034 -0.031 -0.025 -0.057 -0.078 -0.147 -0.218 -0.187 -0.213 -0.106 -0.119 -0.020 0.127 0.617

0.343 0.382 0.186 0.108 0.065 0.018 0.018 -0.121 -0.078 -0.139 -0.168 -0.224 -0.304 -0.267 -0.308 -0.217 -0.241 -0.175 -0.115 0.269

CM 0.403 0.279 0.307 0.083 0.096 0.017 0.077 -0.070 -0.039 -0.196 -0.172 -0.205 -0.244 -0.137 -0.139 -0.183 -0.177 -0.160 0.074 0.372

0.665 0.492 0.285 0.203 0.182 0.130 0.103 0.076 0.030 0.036 0.007 -0.013 -0.009 -0.005 0.044 0.201 0.276 0.532 0.853 2.475

Table 4 Mean Ex post Valuation Errors of Valuation Techniques for Selected Horizons, for Portfolios Formed from a Ranking on Free Cash Flow (to Equity) to Price Portfolio

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Portfolio Attributes at t Mean FCFe /P

Mean GAAP E/P

-1.851 -.505 -.311 -.216 -.153 -.107 -.071 -.042 -.019 -.000 .015 .030 .047 .067 .094 .128 .181 .271 .484 2.697

-.001 .068 .077 .084 .083 .078 .079 .080 .077 .079 .078 .085 .089 .095 .100 .104 .105 .096 .065 -.178

t+1

Horizon t+5

t+8

Price Model -0.016 -0.026 -0.009 -0.016 -0.022 -0.007 -0.030 -0.035 -0.039 -0.028 -0.034 -0.056 -0.067 -0.066 -0.082 -0.098 -0.089 -0.132 -0.137 -0.109

-0.507 -0.337 -0.220 -0.118 -0.143 -0.128 -0.157 -0.139 -0.085 -0.089 -0.142 -0.161 -0.192 -0.231 -0.271 -0.346 -0.422 -0.481 -0.549 -0.751

-0.932 -0.479 -0.306 -0.180 -0.202 -0.256 -0.203 -0.289 -0.163 -0.167 -0.278 -0.252 -0.345 -0.383 -0.417 -0.580 -0.675 -0.746 -0.807 -1.188

Horizon t+5

t+1

t+8

t+1

DCFM (TV: 1.04) 0.948 1.479 2.364 2.500 2.105 1.814 1.654 1.300 1.087 1.081 1.028 0.625 0.719 0.773 0.760 0.598 0.087 -0.023 0.121 -0.596

1.978 2.123 1.603 1.449 1.134 1.122 1.292 0.991 0.913 0.611 -0.174 0.592 0.382 -0.079 -0.207 0.016 0.406 -0.419 0.304 2.243

4.220 2.358 1.222 1.419 1.225 1.040 0.021 0.861 0.606 0.540 0.773 0.281 0.371 0.060 -0.312 0.141 -0.060 0.864 0.967 3.890

Horizon t+5

t+8

RIM -0.049 0.069 0.132 0.184 0.242 0.289 0.322 0.327 0.346 0.384 0.360 0.298 0.220 0.157 0.053 -0.021 -0.130 -0.181 -0.303 -0.231

-0.007 0.051 0.120 0.162 0.193 0.200 0.204 0.220 0.251 0.277 0.251 0.171 0.096 0.010 -0.081 -0.162 -0.250 -0.265 -0.356 -0.320

-0.123 -0.038 0.001 0.078 0.123 0.125 0.114 0.115 0.173 0.170 0.122 0.070 0.002 -0.112 -0.180 -0.270 -0.358 -0.357 -0.431 -0.462

Table 4 (continued) Portfolio Portfolio

Horizon Attributes at tt+1

Mean Mean (Debt+PS)/P GAAP B/P 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

4.164 2.824 2.005 1.540 1.265 1.111 .928 .869 .867 .642 .505 .486 .439 .505 .564 .683 .952 1.227 1.649 3.203

t+5

t+8

RIM (TV: 1.0) 1.147 .979 .908 .845 .760 .716 .672 .668 .653 .614 .626 .667 .720 .791 .880 .945 1.063 1.125 1.289 1.259

0.782 0.488 0.426 0.444 0.269 0.346 0.282 0.221 0.266 0.285 0.270 0.156 0.116 0.066 0.023 -0.011 -0.015 0.024 0.188 0.509

Notes: See notes to Tables 1 and 2.

0.073 0.109 0.221 0.208 0.268 0.188 0.155 0.196 0.278 0.251 0.183 0.125 0.119 -0.062 -0.091 -0.112 -0.109 -0.148 -0.226 -0.069

Horizon t+1

t+5

t+8

0.311 0.167 0.234 0.235 0.197 0.170 0.122 0.162 0.208 0.208 0.160 0.099 0.034 -0.045 -0.098 -0.185 -0.190 -0.161 -0.137 -0.061

-0.001 0.033 0.061 0.131 0.153 0.127 0.097 0.093 0.171 0.130 0.061 0.047 0.002 -0.103 -0.137 -0.226 -0.254 -0.228 -0.227 -0.335

CM 0.343 0.213 0.214 0.164 0.279 0.107 0.168 0.058 0.115 0.090 0.060 -0.081 -0.056 -0.013 -0.266 -0.185 -0.160 -0.146 -0.189 0.053

1.187 0.581 0.444 0.394 0.280 0.324 0.260 0.274 0.291 0.266 0.223 0.175 0.037 0.089 -0.021 -0.050 0.008 0.060 0.358 0.982

Table 5 Mean Ex post Valuation Errors of Valuation Techniques for Selected Horizons, for Portfolios Formed from a Ranking on GAAP E/P Ratios Portfolio

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Portfolio Attributes at t Mean GAAP E/P

Mean FCFe /P

-1.256 -.223 -.055 .003 .027 .042 .055 .062 .072 .079 .087 .094 .102 .111 .119 .130 .144 .162 .195 .358

1.051 .047 -.041 -.040 -.035 -.056 -.040 -.056 -.043 -.040 -.029 -.030 -.022 -.019 -.001 -.015 -.004 -.019 -.035 .148

t+1

Horizon t+5

t+8

Price Model 0.050 0.030 0.062 0.066 0.048 0.046 0.044 0.037 0.022 -0.010 -0.019 -0.035 -0.042 -0.056 -0.076 -0.108 -0.103 -0.128 -0.148 -0.177

-1.163 -0.383 -0.179 0.067 0.057 -0.051 -0.061 -0.057 -0.078 -0.113 -0.166 -0.200 -0.266 -0.277 -0.350 -0.417 -0.441 -0.527 -0.523 -0.653

-1.382 -0.534 -0.152 0.009 -0.001 -0.108 -0.084 -0.146 -0.148 -0.252 -0.284 -0.381 -0.390 -0.478 -0.570 -0.668 -0.676 -0.821 -0.896 -1.091

Horizon t+5

t+1

t+8

t+1

DCFM (TV: 1.04) -4.165 -0.602 0.407 1.082 0.341 0.881 1.185 1.439 1.631 1.362 1.184 1.104 0.928 1.102 0.571 1.120 0.426 0.867 0.862 0.928

3.430 0.846 1.620 0.948 0.812 1.192 0.944 1.584 0.974 0.665 1.079 0.920 0.567 0.599 0.276 -0.158 0.523 -0.281 0.157 -0.058

3.080 1.332 2.193 1.154 0.959 0.621 0.706 0.314 0.731 0.834 0.816 0.981 0.149 0.043 -0.143 -0.121 -0.263 0.313 -0.579 0.941

Horizon t+5

t+8

RIM 0.491 0.063 0.233 0.415 0.426 0.369 0.386 0.361 0.337 0.290 0.252 0.197 0.140 0.072 0.005 -0.080 -0.133 -0.256 -0.348 -0.532

0.126 0.133 0.298 0.429 0.409 0.353 0.317 0.286 0.220 0.165 0.092 0.052 -0.028 -0.069 -0.142 -0.239 -0.268 -0.362 -0.429 -0.608

0.018 0.059 0.223 0.360 0.350 0.271 0.214 0.194 0.107 0.031 -0.040 -0.103 -0.187 -0.201 -0.296 -0.377 -0.395 -0.477 -0.553 -0.711

Table 5 (continued) Portfolio Portfolio

Horizon Attributes at tt+1

Mean Mean (Debt+PS)/P GAAP B/P 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

4.578 1.751 .922 .619 .739 .772 .681 .904 .831 .661 .605 .657 .676 .696 .893 1.069 1.197 1.425 1.781 2.580

t+5

t+8

RIM (TV: 1.0) 1.120 1.095 .764 .588 .596 .635 .631 .647 .663 .691 .700 .732 .783 .848 .895 .957 1.022 1.128 1.211 1.438

1.708 1.304 0.989 0.753 0.722 0.640 0.522 0.434 0.261 0.255 0.186 0.066 0.065 0.009 -0.060 -0.206 -0.149 -0.210 -0.243 -0.207

Notes: See notes to Tables 1 and 2.

0.053 0.302 0.368 0.429 0.369 0.401 0.277 0.259 0.182 0.070 0.082 -0.055 -0.141 -0.111 -0.177 -0.243 -0.274 -0.179 -0.422 -0.434

Horizon t+1

t+5

t+8

0.484 0.534 0.530 0.547 0.425 0.411 0.285 0.270 0.178 0.118 0.019 -0.008 -0.096 -0.114 -0.197 -0.295 -0.284 -0.359 -0.379 -0.487

0.123 0.250 0.312 0.397 0.338 0.266 0.193 0.180 0.098 0.012 -0.047 -0.120 -0.195 -0.201 -0.299 -0.363 -0.359 -0.440 -0.483 -0.577

CM 0.382 0.335 0.503 0.374 0.383 0.202 0.058 0.175 0.117 0.018 0.031 -0.169 -0.118 -0.146 -0.354 -0.348 -0.165 -0.382 -0.498 -0.302

4.215 1.897 1.240 1.095 0.777 0.509 0.319 0.456 0.345 0.286 0.156 0.082 0.052 -0.018 -0.118 -0.241 -0.209 -0.381 -0.495 -0.584

Table 6 Mean Ex post Valuation Errors of Valuation Techniques for Selected Horizons, for Portfolios Formed from a Ranking on the Difference Between GAAP Earnings and Free Cash Flow to Price Portfolio

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Portfolio Attributes at t

Mean (GAAP E-FCFe )/P

Mean FCFe /P

-4.168 -.568 -.249 -.120 -.050 -.009 .018 .039 .058 .075 .093 .114 .139 .169 .206 .255 .321 .421 .617 2.028

2.40881 0.42296 0.24194 0.16030 0.10691 0.07906 0.06130 0.04779 0.03818 0.02719 0.01534 0.00436 -0.01991 -0.04539 -0.08267 -0.12375 -0.17973 -0.26757 -0.44429 -1.68530

t+1

Horizon t+5

t+8

Price Model -0.062 -0.069 -0.074 -0.066 -0.050 -0.045 -0.034 -0.037 -0.038 -0.051 -0.063 -0.061 -0.058 -0.053 -0.059 -0.058 -0.050 -0.046 -0.051 -0.045

-0.926 -0.508 -0.396 -0.350 -0.186 -0.177 -0.128 -0.130 -0.146 -0.177 -0.175 -0.203 -0.171 -0.231 -0.208 -0.205 -0.262 -0.247 -0.406 -0.531

-1.348 -0.764 -0.618 -0.519 -0.289 -0.286 -0.250 -0.224 -0.234 -0.267 -0.285 -0.332 -0.265 -0.396 -0.292 -0.345 -0.389 -0.398 -0.608 -0.971

Horizon t+5

t+1

t+8

t+1

DCFM (TV:1.04) -3.282 0.233 0.287 0.276 -0.232 0.839 0.864 0.633 0.751 0.678 1.069 1.110 1.054 1.289 1.381 1.835 2.014 2.513 2.445 1.202

3.755 0.375 -0.013 0.124 0.221 0.449 0.173 0.437 0.101 0.430 -0.154 1.050 0.883 0.986 1.082 1.452 0.846 2.230 1.256 2.023

5.453 2.079 0.668 0.076 0.365 -0.033 0.424 0.564 0.396 0.163 0.644 0.505 0.737 -0.348 0.779 0.979 0.958 1.340 2.324 4.150

Horizon t+5

t+8

RIM 0.057 -0.217 -0.108 0.018 0.167 0.246 0.311 0.321 0.302 0.299 0.270 0.227 0.194 0.176 0.162 0.122 0.049 0.017 -0.040 -0.156

-0.128 -0.229 -0.133 -0.055 0.078 0.144 0.198 0.203 0.169 0.173 0.139 0.117 0.066 0.082 0.049 0.033 -0.014 -0.018 -0.080 -0.094

-0.289 -0.303 -0.212 -0.136 0.011 0.077 0.125 0.103 0.079 0.066 0.058 0.025 -0.058 -0.021 -0.056 -0.041 -0.137 -0.095 -0.199 -0.196

Table 6 (continued) Portfolio Portfolio Mean GAAP E/P 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Horizon Attributes at tt+1

Mean GAAP B/P -.755 -.080 .024 .058 .067 .077 .080 .083 .086 .091 .096 .099 .102 .102 .105 .108 .112 .114 .125 .127

t+5

t+8

RIM (TV: 1.0) 1.205 1.318 1.180 .982 .828 .749 .683 .653 .646 .664 .700 .731 .766 .790 .812 .862 .922 .984 1.039 1.183

0.956 0.523 0.365 0.299 0.320 0.273 0.268 0.264 0.216 0.198 0.216 0.176 0.183 0.179 0.220 0.215 0.182 0.290 0.309 0.600

Notes: See notes to Tables 1 and 2.

0.331 0.072 0.134 0.167 0.241 0.217 0.212 0.160 0.174 0.173 0.217 0.156 0.128 0.146 0.095 0.164 0.105 0.119 0.078 0.221

Horizon t+1

t+5

t+8

0.123 0.043 0.051 0.016 0.053 0.098 0.127 0.132 0.089 0.098 0.081 0.082 0.012 0.056 0.006 0.038 0.039 0.065 0.022 0.208

-0.266 -0.079 -0.074 -0.073 0.036 0.077 0.101 0.087 0.071 0.049 0.048 0.033 -0.037 -0.015 -0.041 0.006 -0.075 -0.035 -0.138 -0.056

CM 0.526 -0.027 0.082 0.120 0.115 0.121 0.142 0.157 0.148 0.108 0.174 0.113 -0.029 0.104 0.130 0.117 0.023 0.187 0.094 0.324

2.293 0.853 0.438 0.314 0.204 0.274 0.235 0.181 0.152 0.158 0.174 0.102 0.084 0.119 0.145 0.154 0.120 0.212 0.269 0.745

Table 7 Mean Ex post Valuation Errors of Valuation Techniques for Selected Horizons, for Portfolios Formed from a Ranking on B/P Ratios and E/P Ratios Portfolio

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Portfolio Attributes at t Mean GAAP B/P

Mean GAAP E/P

.091 .319 .361 .400 .432 .476 .523 .584 .645 .691 .685 .996 2.215 1.710 1.492 1.449 1.438 1.465 1.516 1.833

-.522 -.012 .028 .045 .057 .069 .081 .094 .107 .126 .192 .108 -.581 -.054 .056 .097 .125 .149 .183 .328

t+1

Horizon t+5

t+8

Price Model 0.091 0.087 0.072 0.061 0.024 0.004 -0.012 -0.029 -0.038 -0.089 -0.118 -0.037 -0.017 0.028 0.002 -0.018 -0.057 -0.100 -0.144 -0.187

-0.471 0.059 0.148 0.083 -0.004 -0.065 -0.101 -0.160 -0.270 -0.404 -0.462 -0.245 -0.670 -0.277 -0.252 -0.239 -0.387 -0.481 -0.528 -0.720

-0.740 0.019 0.107 0.052 -0.068 -0.152 -0.179 -0.249 -0.418 -0.552 -0.719 -0.342 -0.839 -0.338 -0.289 -0.425 -0.667 -0.772 -0.878 -1.091

Horizon t+5

t+1

t+8

t+1

DCFM (TV: 1.04) 0.163 1.103 1.166 1.385 1.553 1.268 0.930 0.923 0.892 0.836 1.026 0.776 -4.410 -0.309 -1.345 0.674 0.032 -0.052 0.476 -0.331

2.001 1.235 1.123 1.049 1.104 1.139 0.727 0.630 0.809 0.582 -0.176 -0.103 1.376 1.728 0.828 0.901 -0.319 -0.026 -0.521 -0.791

2.557 1.576 1.014 0.839 0.548 0.734 0.698 0.465 -0.011 0.653 0.670 -0.324 1.725 1.244 0.813 0.916 0.289 -0.181 -0.108 -0.405

Horizon t+5

t+8

RIM 1.065 0.704 0.649 0.602 0.556 0.498 0.441 0.382 0.304 0.252 0.233 0.022 -0.615 -0.453 -0.360 -0.343 -0.371 -0.437 -0.534 -0.814

0.626 0.651 0.561 0.528 0.442 0.364 0.292 0.227 0.144 0.042 0.004 -0.016 -0.163 -0.175 -0.184 -0.229 -0.310 -0.424 -0.521 -0.695

0.499 0.557 0.489 0.471 0.332 0.241 0.173 0.115 0.020 -0.137 -0.165 -0.083 -0.182 -0.149 -0.200 -0.285 -0.394 -0.504 -0.559 -0.766

Table 7 (continued) Portfolio

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Portfolio Attributes at t Mean (Debt+PS)/P

Mean FCFe /P

1.846 .454 .455 .408 .481 .445 .425 .513 .662 .833 1.099 1.399 3.708 2.232 2.188 1.839 1.833 1.762 2.340 3.694

.457 -.091 -.081 -.079 -.058 -.052 -.043 -.033 -.029 -.028 .014 .032 .641 .061 .029 -.017 .012 .019 -.016 .160

Notes: See notes to Tables 1 and 2.

Horizon t+5

t+1

t+8

RIM (TV: 1.0) 0.917 0.824 0.661 0.580 0.386 0.306 0.188 0.129 0.068 -0.096 -0.266 0.182 2.128 1.161 0.590 0.487 0.150 -0.092 -0.115 -0.041

0.358 0.421 0.504 0.467 0.262 0.209 0.135 0.034 -0.035 -0.244 -0.281 -0.021 0.474 0.376 0.080 0.031 -0.131 -0.247 -0.177 -0.665

0.240 0.604 0.451 0.297 0.267 0.072 0.205 0.037 -0.066 -0.065 -0.177 -0.139 0.428 0.129 -0.022 -0.072 -0.344 -0.364 -0.330 -0.274

Horizon t+5

t+1

t+8

t+1

RIM (TV: 1.04) 0.916 0.753 0.519 0.383 0.099 -0.011 -0.175 -0.264 -0.353 -0.604 -0.892 -0.192 2.689 1.257 0.403 0.269 -0.238 -0.604 -0.642 -0.556

0.056 0.169 0.330 0.266 -0.019 -0.084 -0.170 -0.315 -0.410 -0.711 -0.805 -0.384 0.238 0.162 -0.247 -0.290 -0.527 -0.679 -0.593 -1.412

-0.079 0.489 0.284 0.035 0.050 -0.233 -0.001 -0.242 -0.390 -0.352 -0.562 -0.486 0.183 -0.138 -0.327 -0.370 -0.772 -0.791 -0.775 -0.734

Horizon t+5

t+8

CM 1.644 1.130 0.706 0.479 0.398 0.296 0.190 0.150 -0.041 -0.173 -0.442 0.153 4.355 1.678 0.901 0.580 0.226 -0.056 -0.289 -0.520

0.332 0.661 0.507 0.450 0.327 0.234 0.147 0.062 -0.051 -0.218 -0.318 -0.029 1.010 0.398 0.155 0.051 -0.130 -0.285 -0.373 -0.394

0.247 0.496 0.438 0.419 0.247 0.161 0.097 0.028 -0.101 -0.300 -0.370 -0.104 0.331 0.183 -0.018 -0.150 -0.311 -0.419 -0.450 -0.592